Is Event-Based Prospective Memory Resistant to Proactive Interference?

Transcription

1 DOI /s Is Event-Based Prospective Memory Resistant to Proactive Interference? Joyce M. Oates 1 & Zehra F. Peynircioğlu 1 & Kathryn B. Bates 1 # Springer Science+Business Media New York 2015 Abstract Proactive interference builds up rapidly in recall with the presentation of successive lists comprising items from a single category (e.g., animals). In two experiments, we used a similar build-up paradigm in prospective memory. Interestingly, in Experiment 1, there was no evidence of proactive interference in prospective memory, although build-up emerged in a similar recall task. In Experiment 2, by also showing proactive interference build-up in a recognition task analogous to our prospective memory task, we ruled out the possibility that it was the recognition processes inherent in the prospective memory task that might make the task easier and prevent proactive interference. We suggest that, under normal conditions, prospective memory is resistant to build-up of proactive interference and propose that this resistance is a function of the strengthening afforded by a cue-to-intention association that is a part of prospective memory tasks. We discuss the finding using the classical paired-associate learning framework. Keywords Prospective memory. Proactive interference. Retrospective memory. Proactive inhibition. Build-up of proactive interference Unlike retrospective memory (remembering the past) prospective memory is remembering to carry out a task in the future. Prospective memory is involved in tasks such as remembering to collect your keys before you leave the house, remembering to buy bread after work, or remembering to keep a doctor s appointment 2 months from now. It is often prompted by a specific cue, which is itself an item that needs to be remembered. Thus, prospective memory can require extra monitoring. For instance, to be able to complete the prospective memory task of giving your colleague a message, not only would you need to remember to give the message, but you would * Joyce M. Oates joyce.oates@american.edu 1 Department of Psychology, American University, Washington, DC 20016, USA

2 also need to remember that seeing your colleague was to act as a cue to remind you of this task. In this way, this task could fail for either of two reasons: failing to remember the message or failing to identify seeing your colleague as the appropriate cue to give the message. Although both retrospective and prospective memory are involved in the retrieval of information, the primary difference is that prospective memory relies on self-initiation for the identification of a cue that would trigger the memory for the intention (McDaniel and Einstein 2007). That is, whereas participants are prompted by the experimenter to retrieve the studied items in retrospective memory tasks, there is no such prompting to help participants get into a retrieval mode in prospective memory tasks. Thus, even though, as in all memory, prospective memory also relies on cues, there are two extra steps of 1) noticing the cue, and 2) identifying the cue before the Bcontent^ of the memory can be accessed (Graf 2005). These cues can be Beventbased,^ such as a person or a place, or Btime-based,^ such as time of day or elapsed time (McDaniel and Einstein 2007). To the extent that prospective memory and retrospective memory rely on similar processes, performance in prospective memory should also be vulnerable to retroactive and proactive memory interference. Retroactive interference occurs when newly learned items (e.g., words) interfere with the retrieval of previously learned items whereas proactive interference occurs when previously learned items interfere with the retrieval of newly learned items. Thus, the type of interference that occurs depends on whether previously learned or newly learned items (e.g., words) are tested. Classic interference theory (e.g., Underwood 1957) explains the phenomena as item retrieval competition. Indeed, basic retroactive and proactive interference have already been shown in prospective memory tasks. Although there is some evidence to suggest that retroactive interference in prospective memory may not be as robust as in retrospective memory (Hicks et al. 2000; Martin et al. 2011), Loftus (1971) has demonstrated a healthy retroactive interference in prospective memory. Prior to completing a survey, (ongoing task), participants were instructed to remember to tell the experimenter which state they were born in when the survey was completed (prospective memory task). The variables manipulated were the length of survey (5 or 15 questions) and the presence of a retrieval cue(cue absent just the initial instructions or cue present the initial instructions as well as the information that the last question would be about the Black Panther Party). Loftus predicted that the presence of a retrieval cue and shorter retention duration would aid memory, and indeed found that memory was better when there were fewer intervening questions (i.e., less retroactive interference) and also when a cue was available. Scullin and McDaniel (2010) have also shown retroactive interference in eventbased prospective memory. Further, two studies have reported proactive interference in time-based prospective memory (Cicogna et al. 2005; Occhionero et al. 2010). Interestingly, however, there was no evidence of proactive interference in event-based prospective memory (Cicogna et al. 2005). This was a surprising finding given the time-based prospective memory results, as well as the robustness of proactive interference in general, even in implicit memory (e.g., Lustig and Hasher 2001). One reason Cicogna et al. gave for the lack of proactive interference in their study where the primary task was event-based was that the interpolated prospective memory task was time-based (involving time monitoring), and thus because the two prospective memory

3 tasks were dissimilar, the interpolated task did not induce the expected proactive interference in the primary task. The main purpose of the present study was to test the implied prediction from the Cicogna et al. (2005) study that proactive interference would emerge as in time-based prospective memory if the primary and the interpolated tasks were both event-based. To this end, we used a proactive interference build-up paradigm (Wickens 1970) inwhich performance decreases as the number of successive lists of similar to-be-remembered items (usually same-category items) increases. Of interest was whether prospective memory performance would show a similar decrease, as predicted. Using the same stimuli and design, we tested detection in prospective memory (Experiment 1a) and recall in retrospective memory (Experiment 1b). Each memory task (primary task) was embedded within an ongoing lexical decision task, using words and non-words, where participants decided whether a string of letters was a word (e.g., word) or not (e.g., wurd). In the prospective memory task, the goal was to identify the critical items in lists comprising words from the same category (e.g., animals). We did indeed find a robust proactive interference effect in the recall task. However there was no evidence for any proactive interference in the prospective memory task. One possibility was that in this particular procedure the participants were simply relying on recognition or matching processes for the detection task in prospective memory, which enabled such a task to be resistant to the effects of proactive interference. Although proactive interference build-up has been reported in recognition memory as well (Petrusic and Dillon 1972), most studies in retrospective memory have involved free recall (e.g., Keppel and Underwood 1962; Underwood 1957; Wickens 1970). Thus, possible boundary conditions in recognition paradigms are not well known, and it might be that when relying on processes underlying recognition, proactive interference build-up in all procedures is not as robust as in free recall. To rule out such an explanation in the present case, in Experiment 2, we tested a recognition analogue of our prospective memory task in retrospective memory to make sure that it was not reliance on recognition/matching processes that prevented proactive interference from emerging in the prospective memory experiment. Regardless, the expectation was that proactive interference would build up in the recognition memory task. Experiments 1a and 1b Method Participants Forty American University students (13 females and seven males in Experiment 1a, 15 females and five males in Experiment 1b) participated for extra credit in psychology courses. Materials, Design, and Procedure The words were obtained from Battig and Montague s (1969) category norms and the non-words from the ARC Non-word Database (Rastle et al. 2002). There were 120 words in each of two categories, animals (where Bfour-footed animals,^ Bbirds,^ and Binsects^ categories were combined) and household items (where Bkitchen utensils^ and Bfurniture^ categories were combined). The 120 non-words were all pseudo-homophones (e.g., Btroals^). All words and non-words

4 were 3 10 letters in length with an average length of 5 6 letters and an average number of 2 3 syllables. The number of letters and syllables were not matched between words and non-words. There were three lists, each comprising 30 words and 30 non-words presented twice and in a random order; thus there were 120 items in each list. No item appeared in more than one list. In Experiment 1a, Because proactive interference build-up in retrospective memory has been shown to occur with as few as three items per list (e.g., Wickens 1970), and because it has also been shown in recognition using only six items per list (i.e., three targets and three foils e.g., Petrusic and Dillon 1972), we chose three of the words (e.g., lizard, zebra, coyote) to serve as prospective memory cues in each list. In order to keep level of prospective memory cue difficulty equal across categories, only words of 5 6 letters in length were used. These cues were repeated once to make six prospective memory trials, and were presented randomly among the other stimuli (also repeated once each) but within the constraint that each cue was spaced apart by at least 15 other items. Each cue was assigned to one of three colored (red, green, or yellow) button responses at the beginning of each list. The three build-up lists comprised words (and cues) all from the same category (animals or household items). Build-up lists were counterbalanced across participants such that half the participants had build-up lists from the household items category, and the other half had build-up lists from the animals category. The prospective memory task was embedded in a lexical decision task; a string of letters appeared in the middle of the computer screen and remained until a word or nonword judgment was made as quickly and accurately as possible. Participants were told at the beginning of the experiment that the primary task was making word judgments. In addition, they would need to keep a few words in mind (prospective memory cues) and indicate the appearance of these words by making a special button press (e.g., press the Bred^ button if Blizard^ appears). Repetition of the cues ensured that participants could not just forget about the word after its first presentation. That is, participants had to keep the cues in mind throughout each entire list. It was stressed by the experimenter that the special button press would have to be made after the lexical decision response. New prospective memory cues were assigned at the beginning of each list. Participants were required to remember only the prospective memory cues. See the Appendix for an example of the task instructions 1. To clarify the instructions, participants first completed a short practice list. Participants were tested individually on a computer, and the experiment script was coded in SuperLab 4.0 using a Cedrus RB-830 button-box to collect all button presses and the reaction times associated with those button presses. In each trial, a preparatory fixation cross B+^ appeared in the middle of the computer screen for 250 ms, after which a string of letters appeared in the center of the screen. Participants pressed the BW^ button for a word or the BN^ button for a non-word decision. Immediately after the lexical decision, the italicized word Waiting appeared in the center and the participants pressed either BW^ or BN^ again to advance to the next trial regardless of what key they had pressed in response to the stimulus itself. If the item also happened to be a prospective memory cue, participants pressed the colored button 1 Although the onscreen instructions only made reference to working as accurately as possible, the experimenter explained verbally that it was also important to work as quickly as possible without losing accuracy.

5 associated with that cue during the Waiting screen, before pressing BW^ or BN^ again to advance to the next trial (cf. Marsh et al. 2002). Proportion correct for the task was dependent upon making a correct color button press. Prospective memory reaction times were the latencies to make the colored button presses. In Experiment 1b, the materials and design were the same as in Experiment 1a, and the procedure was also similar in that participants made lexical decisions; however, instead of doing a prospective memory task, participants were told before beginning the experiment that they would have to remember the words (not the non-words) and received a recall test after each list. This task also included the prospective memory cue words used in Experiment 1a, although participants were unaware that those particular words had been used as prospective memory cues in a different experiment. For this test, they were given a sheet of paper and pen and instructed to write down as many of the words as they could remember from the last list. They were given 2 minutes and timed with a stop-watch. Results and Conclusions The results of Experiment 1a are summarized in Table 1. We conducted linear contrast ANOVAs with build-up lists in order to test for a trend of proactive interference buildup. There was no main effect of build-up list in proportion correct, F<1. Prospective memory accuracy did not decrease across lists. Further, prospective memory response reaction times decreased across lists, F(1, 19)=24.5, p<.05, MSE= , partial η 2 =.56, showing that participants were not trading accuracy for an increase in speed. Conversely, in Experiment 1b, the typical proactive interference build-up pattern emerged. Across lists, proportion correctly recalled decreased,.40 (.13),.35 (.11),.32 (.09), respectively, F(1, 19)=5.0, p<.05, MSE=.05, partial η 2 =.21, indicating a buildup of proactive interference. One explanation for the lack of proactive interference build-up in Experiment 1a could be that participants were allotting increasingly greater cognitive resources away from the ongoing task in order to stave off proactive interference in the prospective memory task refer to Table 1 for the descriptive statistics. However, accuracy in the lexical decision task did not decrease across lists, F(1, 19)=1.5, p>.10, MSE=.001, partial η 2 =.07, and, similar to prospective memory responses, reaction times decreased across lists, F(1, 19)=23.8, p<.05, MSE=150860, partial η 2 =.56. Thus, it was unlikely that the lack of a proactive interference build-up was the result of reallocating more attention to the prospective memory task at the expense of the ongoing task. Table 1 Experiment 1a: Proportions correct and reaction times (RTs) for prospective memory (ProM) and lexical decision (LD) tasks as a function of list List ProM correct ProM RTs LD correct LD RTs (.23) 1913 (589) 0.96 (.05) 851 (137) (.21) 1601 (459) 0.94 (.03) 772 (115) (.11) 1328 (286) 0.97 (.03) 728 (98) The standard deviations are in parentheses

6 There was no proactive interference build-up in prospective memory. One possible explanation for this lack of an effect might be that the prospective memory task did not involve effortful retention of the target items and could be performed simply by reacting to the cue when it appeared. To rule out this explanation, in Experiment 2, we tested for a proactive interference build-up in a retrospective recognition paradigm that was analogous to our prospective memory experiment and used exactly the same stimuli and ongoing task (lexical decision). Experiment 2 Method Participants Twenty American University students (18 women and two men) participated for extra credit in psychology courses. None had participated in the previous experiments. Materials, Design, and Procedure The materials and design were the same as in Experiments 1a and 1b. The procedure differed in that the measure of interest was participants retrospective recognition performance. As in Experiments 1a and 1b, after a short practice session, participants made a series of lexical decisions in three different lists. Unlike in Experiments 1a and 1b, however, each list presentation was followed by a recognition test. Further, nine of the words that would be scattered among the three ensuing list presentations were shown at the start of the experiment with the instructions for the participants to keep these target words in mind for the recognition tests. Indeed, three of the nine target words presented at the start of the experiment appeared in each lexical decision list, and thus the recognition test for each list consisted of three to-beremembered target words and nine lures. No special tasks were associated with these target words during list presentation. That is, while making the lexical decisions, there was no additional task of indicating whether or not a given stimulus was one of the target words to keep in mind. The purpose of presenting these words in the beginning was simply to give participants this additional mental load in keeping with the procedure of the Experiment 1b. During each recognition test, 12 words were presented one at a time on the computer screen. All of these words had appeared during the preceding list presentation, but only three were also from the original nine that the participants had been asked to keep in mind and served as the targets while the remaining words, intermixed randomly, served as the lures. Each word remained on the screen until the participant indicated whether or not that word was one of the initial nine target words by pressing the green button for a BYes^ response and the red button for BNo^ response. None of these test words appeared again in another presentation list or recognition test. Results To test for proactive interference build-up in recognition performance, signal detection analyses were conducted. The d prime values were 1.87 (.76), 1.66 (1.20), and 1.01

7 (1.19), for Lists 1, 2, and 3, respectively, and a linear contrast repeated measures ANOVA showed this decreasing trend to be significant, F(1,19)=10.6, p<.05, MSE= 7.4, partial η 2 =.36. Thus, proactive interference did build up in retrospective recognition memory with the same materials and design as in our prospective memory paradigm. Discussion Loftus (1971) had made the prediction that because prospective memory involved similar processes as retrospective memory, traditional interference effects should be observed in prospective memory tasks, as well. Indeed, she and others (e.g., Scullin and McDaniel 2010) found comparable effects of retroactive interference on prospective memory in that performance for the prospective memory trial was hurt when the number of interpolated items in the ongoing task increased. Further, Cicogna et al. (2005) found proactive interference in time-based prospective memory. However, they did not find it in event-based prospective memory, and attributed this primarily to the lack of processing similarity between the interpolated and primary tasks (event-based interpolated task but time-based primary task). The assumption then was that when the primary and interpolated tasks are similar (both time-based or both event-based) proactive interference should arise in all prospective memory tasks, just like in retrospective memory tasks. We tested this assumption directly. Even though both our primary and interpolated tasks were event-based and relied on the same items, there was no proactive interference in prospective memory. Thus, lack of similarity of the interpolated task to the primary task was unlikely to have led to the lack of proactive interference in the eventbased prospective memory task of Cicogna et al. (2005). Further, similar procedures with both free recall and recognition did show a proactive interference build-up, suggesting that the lack of proactive interference in prospective memory was also not due to our specific methodology or to lack of effortful retention of the prospective memory items. That is, this conceptual replication of the classic build-up of proactive interference paradigm showed that prospective memory was different from retrospective memory in that whereas both recall and recognition (presumably less effortful than recall and more akin to prospective memory) showed build-up of proactive interference, prospective memory did not. One possible reason for why proactive interference build-up was not observed in prospective memory is because the present paradigm involved, for all practical purposes, paired-associate learning. This is true in event-based prospective memory tasks by definition, where an event is coupled with and acts as a cue for another event, whereas the cue of time in time-based prospective memory tasks is perhaps more diffuse and does not act as a strict paired associate. In the present case, in classical interference terms, the cue-intention pairings were of the A-B, C-D type (e.g., zebra-red button, lizard-yellow button, etc.), and although responses were re-paired with different cues (e.g., monkey-red button), cues were unique and never re-paired with responses that is, Bzebra^ was only paired with the red button and only appeared as a cue on a single list. In neither case would one expect much interference when the cue terms are presented to the participants (e.g., Osgood 1949). Although the theory is primarily for

8 explaining retroactive interference rather than proactive interference, to the extent that the underlying mechanism of interference is similar, these predictions should also hold for proactive interference build up (cf. Dallett 1962). Thus, we would not expect much of an interference effect in such a paradigm. When one needs to relay a message to a specific colleague, (or press a button for Bzebra^), one should not confuse the colleague (cue, Bzebra^) to relay the message to (press the red button), despite the fact that many people encountered during the day will all be from the same category of Bcolleague.^ Interference would likely have emerged had we used A B, A D pairs (where cues such as zebra had been re-paired with several intentions such as red button, yellow button). However, such a finding would apply only to the retrospective memory part of the prospective memory task remembering the intentions associated with the cues and would not be surprising given that similar fan effects have been well-documented in the retrospective memory literature (e.g., Anderson 1974). In other words, the prospective memory component would remain intact and the cue could be identified. This would be akin to intending to relay several messages to a single colleague at different encounters. The cue (colleague) can be easily identified, but the difficulty arises in trying to retrieve the specific message. A caveat is that in classical interference studies, the cues are provided and the task is to remember the responses that go with them. In such cases proactive interference does not occur in retrospective memory. Both Murdock (1964) and Poppei et al. (1970) have demonstrated that, when the stimuli are paired associates and the response is paired with an explicit cue, there is no proactive interference build-up. The present case was not quite analogous, however, because participants were expected to Brecognize^ the cues of their own accord in order to be able to attempt the associated response. As such, one would still expect a build-up of proactive interference given that it has been demonstrated when using recognition tests, as well (e.g., Petrusic and Dillon 1972), and we also demonstrated it in a recognition memory analogue of our prospective memory paradigm. Thus, the question remains: What makes the recognition of the cues as to-be-identified items resistant to proactive interference? Lastly, the following limitations of the present study need to be considered. As in all studies that utilize young, undergraduate samples, there is a caveat with regards to whether or not the results might generalize to the whole population of young adults. Likewise, since prospective memory tends to decline with age, these effects might be different for older populations; older adults may show proactive interference build-up whereas younger adults are resistant to it. We should also mention that word length and syllables were not controlled for or varied systematically, which might have influenced the results. In addition, perhaps proactive interference in prospective memory builds at a slower rate than retrospective memory; adding more lists might allow for build-up. It could also be the case that build-up for prospective memory requires more cues per list, so doubling the number of cues per list might create build-up. These issues are important future considerations in order to determine the nature of the resistance of prospective memory to proactive interference build-up. We suggest that these paired-associate prospective memory cues have a privileged status in that not only do they receive additional processing which makes them more strongly organized (cf. Charness 1976), but they are also made relatively more distinctive compared to other items of the same category and thus less easily interfered with.

9 In that way, they can benefit from a type of Bpop out^ effect. Further, there is evidence in the literature that prospective memory benefits from an intention superiority effect (e.g., Cohen et al. 2011; Goschke and Kuhl 1993; Penningroth et al. 2012). In conclusion, we suggest that in event-based prospective memory, cue-to-intention associations have a privileged status that enables memory to stay resistant to build-up. Acknowledgments Many thanks to Daniel Dickison for his programming help in data analysis. Communication concerning this article can be directed to either Joyce Oates, joyce.oates@american.edu, or Zehra Peynircioğlu, peynir@american.edu. Appendix. Example of the prospective memory task instructions for Experiment 1a You will see letter strings presented one at a time on the computer screen. Your task is to indicate if the string presented is a word or is a non-word by pressing the BN^ (nonword) or BW^ (word) button. For example, if you saw Bword^, you would press BW^ and if you saw Bwurd^, you would press BN^. Please respond as accurately as you can for the word decision task. Some items may appear more than once. In addition you will need to keep in mind 3 words, LIZARD, ZEBRA, and COYOTE. During the BWaiting ^ screen following the word judgment response, press the: red button if the word was LIZARD green button if the word was ZEBRA yellow button if the word was COYOTE Finally, press either BN^ or BW^ to advance to the next trial. References Anderson, J. R. (1974). Retrieval of propositional information from long-term memory. Cognitive Psychology, 6(4), Battig, W. F., & Montague, W. E. (1969). Category norms of verbal items in 56 categories A replication and extension of the Connecticut category norms. Journal of Experimental Psychology, 80(No. 3, Part 2), Charness, N. (1976). Memory for chess positions: resistance to interference. Journal of Experimental Psychology: Human Learning and Memory, 2(6), Cicogna, P., Nigro, G., Occhionero, M., & Esposito, M. (2005). Time-based prospective remembering: interference and facilitation in a dual task. European Journal of Cognitive Psychology, 17(2), Cohen, A. L., Kantner, J., Dixon, R. A., & Lindsay, D. S. (2011). The intention interference effect. Experimental Psychology (formerly Zeitschrift für Experimentelle Psychologie), 58(6), Dallett, K. M. (1962). The transfer surface re-examined. Journal of Verbal Learning and Verbal Behavior, 1(2), Goschke, T., & Kuhl, J. (1993). Representation of intentions: persisting activation in memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 19(5), Graf, P. (2005). Prospective memory retrieval revisited. In Dynamic cognitive processes (pp ). Tokyo: Springer.

10 Hicks, J. L., Marsh, R. L., & Russell, E. J. (2000). The properties of retention intervals and their affect on retaining prospective memories. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26(5), Keppel, G., & Underwood, B. J. (1962). Proactive inhibition in short-term retention of single items. Journal of Verbal Learning and Verbal Behavior, 1(3), Loftus, E. F. (1971). Memory for intentions: the effect of presence of a cue and interpolated activity. Psychonomic Science, 23(4), Lustig, C., & Hasher, L. (2001). Implicit memory is vulnerable to proactive interference. Psychological Science, 12(5), Marsh, R. L., Hicks, J. L., & Watson, V. (2002). The dynamics of intention retrieval and coordination of action in event-based prospective memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 28(4), Martin, B. A., Brown, N. L., & Hicks, J. L. (2011). Ongoing task delays affect prospective memory more powerfully than filler task delays. Canadian Journal of Experimental Psychology/Revue canadienne de psychologie expérimentale, 65(1), McDaniel, M., & Einstein, G. (2007). Prospective memory: An overview and synthesis of an emerging field. Thousand Oaks: Sage Publications, Inc. Murdock, B. B., Jr. (1964). Proactive inhibition in short-term memory. Journal of Experimental Psychology, 68(2), Occhionero, M., Esposito, M. J., Cicogna, P. C., & Nigro, G. (2010). The effects of ongoing activity on time estimation in prospective remembering. Applied Cognitive Psychology, 24(6), Osgood, C. E. (1949). The similarity paradox in human learning: a resolution. Psychological Review, 56(3), Penningroth, S. L., Graf, P., & Gray, J. M. (2012). The effect of a working memory load on the intentionsuperiority effect: examining three features of automaticity. Applied Cognitive Psychology, 26(3), Petrusic, W. M., & Dillon, R. F. (1972). Proactive interference in short-term recognition and recall memory. Journal of Experimental Psychology, 95(2), Poppei, J. E., Finlay, B. L., & Tedford, W. H., Jr. (1970). Proactive inhibition in short-term memory. Journal of Experimental Psychology, 83(1), Rastle, K., Harrington, J., & Coltheart, M. (2002). 358,534 nonwords: the ARC nonword database. The Quarterly Journal of Experimental Psychology Section A, 55(4), Scullin, M., & McDaniel, M. (2010). Remembering to execute a goal: sleep on it! Psychological Science, 21(7), Underwood, B. (1957). Interference and forgetting. Psychological Review, 64(1), Wickens, D. D. (1970). Encoding categories of words: an empirical approach to meaning. Psychological Review, 77(1), 1 15.